Developed GPT-4 + spaCy pipelines to extract and summarize insights from 2K+ global-health reports, accelerating literature review and policy analysis. Built predictive ML models (XGBoost, logistic regression) for maternal and child health risk mapping across 10+ regions. Automated ETL workflows with Airflow, FastAPI, and SQL, cutting data refresh latency by 40%.

Optimized REST APIs (Node.js + Express) for user auth and content delivery — 35% faster average response time serving 10K+ daily requests. Redesigned MySQL schema with indexing and normalization — 40% faster queries on analytics endpoints. Deployed AWS EC2 + RDS backend stack with CloudWatch monitoring for API health.

Built a React + Node.js dashboard for 500+ customers, enabling self-service analytics and reducing ticket volume by 40%. Indexed 150K+ MongoDB documents → reduced dashboard load time 3.5s → 0.8s. Secured APIs with JWT authentication and bcrypt password hashing.

Developed spam detection model using TF-IDF + logistic regression — 85% precision / 82% recall on 3K+ customer tickets. Analyzed 200+ misclassified samples, introducing multilingual and short-query augmentation. Deployed classifier to production, reducing manual review workload by 30% weekly.